A robust machine learning algorithm to search for continuous gravitational waves. (arXiv:2007.08207v1 [astro-ph.IM])
<a href="http://arxiv.org/find/astro-ph/1/au:+Bayley_J/0/1/0/all/0/1">Joseph Bayley</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Messenger_C/0/1/0/all/0/1">Chris Messenger</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Woan_G/0/1/0/all/0/1">Graham Woan</a>
Many continuous gravitational wave searches are affected by instrumental
spectral lines that could be confused with a continuous astrophysical signal.
Several techniques have been developed to limit the effect of these lines by
penalising signals that appear in only a single detector. We have developed a
general method, using a convolutional neural network, to reduce the impact of
instrumental artefacts on searches that use the SOAP algorithm. The method can
identify features in corresponding frequency bands of each detector and
classify these bands as containing a signal, an instrumental line, or noise. We
tested the method against four different data-sets: Gaussian noise with time
gaps, data from the final run of Initial LIGO (S6) with signals added, the
reference S6 mock data challenge data set and signals injected into data from
the second advanced LIGO observing run (O2). Using the S6 mock data challenge
data set and at a 1% false alarm probability we showed that at 95% efficiency a
fully-automated SOAP search has a sensitivity corresponding to a coherent
signal-to-noise ratio of 110, equivalent to a sensitivity depth of 10
Hz$^{-1/2}$, making this automated search competitive with other searches
requiring significantly more computing resources and human intervention.
Many continuous gravitational wave searches are affected by instrumental
spectral lines that could be confused with a continuous astrophysical signal.
Several techniques have been developed to limit the effect of these lines by
penalising signals that appear in only a single detector. We have developed a
general method, using a convolutional neural network, to reduce the impact of
instrumental artefacts on searches that use the SOAP algorithm. The method can
identify features in corresponding frequency bands of each detector and
classify these bands as containing a signal, an instrumental line, or noise. We
tested the method against four different data-sets: Gaussian noise with time
gaps, data from the final run of Initial LIGO (S6) with signals added, the
reference S6 mock data challenge data set and signals injected into data from
the second advanced LIGO observing run (O2). Using the S6 mock data challenge
data set and at a 1% false alarm probability we showed that at 95% efficiency a
fully-automated SOAP search has a sensitivity corresponding to a coherent
signal-to-noise ratio of 110, equivalent to a sensitivity depth of 10
Hz$^{-1/2}$, making this automated search competitive with other searches
requiring significantly more computing resources and human intervention.
http://arxiv.org/icons/sfx.gif
